Performance Analysis of Electrical Machines Using a Hybrid Data- and Physics-Driven Model

In the design phase of an electrical machine, finite element (FE) simulations are commonly used to numerically optimize the performance. The output of the FE simulation is used to characterize the electromagnetic behavior of the machine. The simulation workflow involves intermediate measures such as nonlinear iron losses, electromagnetic torque, and flux values at each operating point that are used to compute several key performance indicators (KPIs). We present a data-driven deep learning approach that replaces the computationally heavy FE calculations by a deep neural network (DNN). The DNN is trained by a large volume of stored FE data in a supervised manner. During the learning process, the network response (intermediate measures) is fed as input to a physics-based post-processing to estimate characteristic maps and KPIs. We show that this hybrid approach reduces the computational time while maintaining a high flexibility in the simulation workflow. Finally, the hybrid approach is compared to the existing direct prediction approach for KPIs.